EH Eric Ho On The MAD Podcast with Matt Turck

“One of our researchers found that the model was explicitly reasoning about having an external chain of thought monitor monitoring its reward hacks. And so the model was reasoning: hey, all right, somebody's going to read my reasoning. So I should think about how to craft my reward hack so it won't get caught by this external monitor.”

The MAD Podcast with Matt Turck · AI Research & Frontier Labs · October 2026

“One of our researchers found that the model was explicitly reasoning about having an external chain of thought monitor monitoring its reward hacks. And so the model was reasoning: hey, all right, somebody's going to read my reasoning. So I should think about how to craft my reward hack so it won't get caught by this external monitor.” — Eric Ho, The MAD Podcast with Matt Turck

Context

The host asks whether models ever hide what they are doing when they cheat on a task. Ho describes a finding from "just a couple of days ago" and says the layers of gaming have become deep enough that he can no longer follow them.

Transcript

The MAD Podcast with Matt Turck Around 26:13 into the episode
Eric Ho

…term right now. I think people are still wrapping their head around what neuralese is. And so, yeah, like those two ideas, like this prevalence of neural lease, means that you're going to have to understand neurons and what they do and what's going on. So that's the whole point of interpretability.

Matt Turck

And are there examples, perhaps as part of your research, where the models intentionally make the chain of thought impossible to understand as part of a reward of hacking, basically sort of hiding their actions?

Eric Ho

That is a really interesting question. We have now, I believe, seen, yeah, so just a couple of days ago, one of our researchers found that the model was explicitly reasoning about having an external chain of thought monitor monitoring. It's reward hacks. And so the model was reasoning: hey, all right, somebody's going to read my reasoning. So I should think about how to craft my reward hack so it won't get caught by this external monitor. And so the meta games with these models of gaming these evaluations are getting so many layers deep that I can't even follow it anymore. You know, it's like, it's pretty complicated. And I think it's just going to be a little bit of a cat and mouse game where the model is just going to try to solve the problems and find the answer no matter what. And sometimes cheating is the easiest way and it's effective.

Matt Turck

And we mentioned the labs and we had mentioned open source models at the beginning of this conversation, but I just want to double-click on the point because it's pretty essential. What we're saying here is that all of this is a problem across open source, closed source. So obviously the closed…

Eric Ho

Yeah, well, actually, my mental model of this is that the vast majority of the world hasn't really seen truly capable and misaligned models yet in training and inference. It's a frontier problem mostly. The open weight models aren't quite capable enough to really pull off and chain complex cyber…

Matt Turck

Great. All right. So that's the safety and alignment, the way it's been done, observability. Let's go back to Interp in more detail. So what's the high-level state of the art, I guess, in Interp? What is it that we know about how those models work? When a model answers a question, what is actually…

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.

More from The MAD Podcast with Matt Turck